Near Real Time Automatic Extraction of CME Properties From LASCO C2 Coronagraph Images Using Deep Learning

TransCME is a automated Python pipeline for downloading SOHO/LASCO C2 images, running CME segmentation, and saving tracking results. seeds_monthly_2023_2025_merged_process.txt is the CME catalog text produced by the SEEDS method. ours.txt is the CME catalog generated by our method. univ_all_CDAW_process.txt is the CDAW catalog, and comb_qkl_CACTUS_process.txt is the catalog produced by the CACTus method. CDAW_matched_1.txt contains the catalog entries from CDAW that were matched with our detections, and ours_cdaw_matched_1.txt contains the corresponding entries from our catalog that match CDAW. public_data_1.tar contains the processing outputs of our method for CMEs from 2023 to 2025, including segmentation, tracking, linear speed fitting, and the final exported catalog.